CD109 drives pro-tumorigenic cell properties in human non-small cell lung cancer through interaction with desmoglein-2.

Cluster of differentiation 109 (CD109) is a glycosylphosphatidylinositol (GPI) anchored cell surface protein, expressed on epithelial and endothelial cells, CD4+ and CD8+ T-cells, and premature lymphocytes. CD109 interacts with different cell surface receptors and thereby modulates intracellular signaling pathways, which ultimately changes cellular functions. One well-studied example is the interaction of CD109 with the TGFß/TGFß-receptor complex at the cell surface. CD109 silences intracellular SMAD2/3 signaling and targets TGFß/TGFß-receptor to the endosomal/lysosomal compartment. In recent years, CD109 emerged as a tumor marker for different tumor entities and expression of CD109 could be linked to adverse outcome in patients. In this study, we show that silencing of CD109 in human non-small cell lung cancer (NSCLC) cells, returns these cells to an epithelial like growth phenotype. On the transcriptional level, we describe changes in cell-cell contact and epithelial-mesenchymal transition associated gene clusters. At the cell surface, we identify desmoglein-2 (DSG2) as a new interaction partner of CD109 and demonstrate CD109 dependent targeting of DSG2 to the apical cell surface, where it forms desmosomes between apical and basal cell poles. Both, CD109 and DSG2 are genetic risk factors, linked to reduced overall survival in lung adenocarcinoma patients (subtype of NSCLC). In this study, we show the expression of both proteins in the same tumor and suggest a new CD109-DSG2 axis in NSCLC patients that could present a targetable therapeutic option in the future.

Disseminated cancer cells detected by immunocytology in lymph nodes of NSCLC patients are highly prognostic and undergo parallel molecular evolution.

In melanoma, immunocytology (IC) after sentinel lymph node disaggregation not only enables better quantification of disseminated cancer cells (DCCs) than routine histopathology (HP) but also provides a unique opportunity to detect, isolate, and analyse these earliest harbingers of metachronous metastasis. Here, we explored lymph node IC in non-small cell lung cancer (NSCLC). For 122 NSCLC patients, 220 lymph nodes (LNs) were split in half and prepared for IC and HP. When both methods were compared, IC identified 22% positive patients as opposed to 4.5% by HP, revealing a much higher sensitivity of IC (p < 0.001). Assessment of all available 2,952 LNs of the same patients by HP uncovered additional patients escaping detection of lymphatic tumour spread by IC alone, consistent with the concept of skip metastasis. A combined lymph node status of IC and complete HP on a larger cohort of patients outperformed all risk factors in multivariable analysis for prognosis (p < 0.001; RR = 2.290; CI 1.407-3.728). Moreover, isolation of DCCs and single-cell molecular characterization revealed that (1) LN-DCCs differ from primary tumours in terms of copy number alterations and selected mutations and (2) critical alterations are acquired during colony formation within LNs. We conclude that LN-IC in NSCLC patients when combined with HP improves diagnostic precision, has the potential to reduce total workload, and facilitates molecular characterization of lymphatically spread cancer cells, which may become key for the selection and development of novel systemic therapies. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

EpCAM-positive disseminated cancer cells in bone marrow impact on survival of early-stage NSCLC patients.

INTRODUCTION: Detection of disseminated cancer cells (DCC) in bone marrow (BM) of patients with early-stage NSCLC has been associated with poor outcome. However, the phenotype, and hence relevant therapy targets, of DCCs in BM are unknown. We therefore compared a classical pan-Cytokeratin (CK) antibody for DCC detection with an anti-EpCAM antibody that may also detect more stem-like cells and tested whether assay positivity impacts on the survival of NSCLC patients. MATERIALS AND METHODS: We prospectively collected BM aspirates from 104 non-metastasized NSCLC patients that underwent potentially curative tumor resection from 2011 to 2016 at the Department of Thoracic Surgery of the University Hospital and Hospital Barmherzige Brüder in Regensburg. DCCs were detected by staining with the pan anti-CK antibody A45-B/B3 and the anti-EpCAM antibody HEA-125. We analyzed the association between detection of DCCs and clinicopathological characteristic and patient outcome. RESULTS: CK + and EpCAM + DCCs were detected in 45.2% and 52.9% of patients, respectively. Correlation between the two markers was low and neither of them was associated with sex, age, histology, T or N classification, resection status, grading or smoking habit. No significant association with tumor specific survival (TSS) and progression-free survival (PFS) was observed in patients with CK + DCCs. In contrast, detection of EpCAM + DCCs significantly correlated with reduced PFS (P = 0.017) and TSS (P = 0.017) and remained an independent prognostic variable for PFS and TSS upon multivariate testing (hazard ratio: 7.506 and 3.551, respectively). Detection of EpCAM + DCCs was the only prognostic marker for PFS. CONCLUSIONS: EpCAM+, but not CK + DCCs in BM predict reduced PFS and TSS. This finding suggests that EpCAM + DCCs in the BM comprise metastatic founder cells necessitating their in-depth molecular analysis for detection of novel therapy targets.